VEHICLE COUNTING DEVICE AND VEHICLE COUNTING METHOD
A vehicle counting device including: a frequency analysis unit that analyzes a frequency of a surrounding sound detected by a vehicle sound detection microphone; a vehicle sound candidate selection unit that selects, as one or more vehicle sound candidates, one or more sounds included in the surrounding sound, based on the analysis by the frequency analysis unit, each of the one or more sounds being in a frequency band where a sound pressure is greater than or equal to a predetermined threshold value; a phase curve calculation unit that calculates, for each of the one or more vehicle sound candidates, a phase curve; and a vehicle count determination unit that classifies the one or more phase curves into at least one group, based on shapes of the one or more phase curves, and determines a total number of the at least one group as the number of the nearby vehicles.
This is a continuation application of PCT Patent Application No. PCT/JP2011/007357 filed on Dec. 28, 2011, designating the United States of America, which is based on and claims priority of Japanese Patent Application No. 2011-004371 filed on Jan. 12, 2011. The entire disclosures of the above-identified applications, including the specifications, drawings and claims are incorporated herein by reference in their entirety.
TECHNICAL FIELDOne or more exemplary embodiments disclosed herein relate generally to devices that determine the number of vehicles, and particularly to a device that detects the number of vehicles, using one or more vehicle sounds from at least one other vehicle around a user vehicle.
BACKGROUND ARTExamples of a conventional technique of determining a situation of a vehicle around a user vehicle include the following techniques.
In the first conventional technique, a surrounding sound is converted into a sound pressure level signal. An absolute quantity of the sound pressure level signal in a specific frequency band is compared with a determination level. It is determined, using the comparison result, whether or not a vehicle is present around a user vehicle. In addition, it is determined whether or not another vehicle is approaching, based on a time dependency of the sound pressure level signal (see Patent Literature (PTL) 1, for instance).
CITATION LISTPatent Literature
[PTL 1]
- Japanese Unexamined Patent Application Publication No. 2000-99853
[PTL 2]
- Japanese Patent No. 2806048
Non Patent Literature
[NPL 1]
Approaching vehicles detection system by using vehicle noise for driver support (Doraiba Shien No Tame No Soko-on Ni Yoru Sekkin Sharyo Kenchi Shisutemu). The Journal of the Acoustical Society of Japan vol. 62, no. 3 (2006): 265-274
SUMMARY OF INVENTION Technical ProblemIn the first conventional technique (PTL 1 and NPL 1), a surrounding sound in a frequency band such as 1000 Hz is converted into a sound pressure level signal, an absolute quantity of the sound pressure level signal in the specific frequency band is compared with a determination level so as to determine whether or not a nearby vehicle is present and determine whether or not the nearby vehicle is approaching, based on a time dependency of the sound pressure level signal (see FIG. 3 of PTL 1, for example). Although, however, it is possible to detect the approaching vehicle, because the high frequency band such as 1000 Hz includes a friction sound between tires and a road surface (a sound of vehicle tires) that is produced by a moving vehicle, when vehicles are present, it is not possible to determine the number of the vehicles. This is because the tire moving sound extends over a wide range such as a range from 800 Hz to 1200 Hz and has no characteristics for use in distinction depending on vehicles. In other words, whether one or two vehicles are present, sound pressure appears in a similar frequency band, and thus not only is it impossible to determine the number of vehicles, but it is difficult to detect more than one vehicle (see NPL 1, page 271).
The second conventional technique (PTL 2) relates to distinguishing, using harmonic information, timbres produced by different musical instruments. Each of the musical instruments produces a sound in a unique frequency band. In addition, the frequency band has a so-called harmonic structure having a fundamental frequency, a frequency twice as high as the fundamental frequency, and so on. Furthermore, sounds in the frequency band are always produced simultaneously. Thus, assuming that one musical instrument produces a frequency having certain sound pressure simultaneously generated and a frequency twice as high as the frequency, each of musical instruments is classified into a group, and a music score is created for each musical instrument.
However, a sound emanating from a vehicle, especially an engine sound, has a unique timbre depending on a vehicle, and is not limited to be a harmonic. Moreover, an SN ratio is small due to a noise such as a noise of the wind or influence such as reflection and diffraction caused by a building or the like in the actual environment, and thus it is difficult to distinguish vehicles by simply using sound pressure information of a specific frequency.
On non-limiting and exemplary embodiment provides a vehicle counting device which accurately determines the number of vehicles, using vehicle sounds from other vehicles around a user vehicle.
Solution to ProblemIn one general aspect, the techniques disclosed here feature a vehicle counting device that determines the number of nearby vehicles, using one or more vehicle sounds, the vehicle counting device including: a frequency analysis unit that analyzes a frequency of a surrounding sound that includes the one or more vehicle sounds and is detected by a vehicle sound detection microphone; a vehicle sound candidate selection unit that selects, as one or more vehicle sound candidates, one or more sounds included in the surrounding sound, based on a result of the analysis by the frequency analysis unit, each of the one or more sounds being in a frequency band where a sound pressure is greater than or equal to a predetermined threshold value; a phase curve calculation unit that calculates, for each of the selected one or more vehicle sound candidates, a phase curve indicating a time dependency of a phase; and a vehicle count determination unit that classifies the calculated one or more phase curves into at least one group, based on shapes of the calculated one or more phase curves, and determines a total number of the at least one group as the number of the nearby vehicles.
These general and specific aspects may be implemented using a system, a method, an integrated circuit, a computer program, or a computer-readable recording medium such as a CD-ROM, or any combination of systems, methods, integrated circuits, computer programs, or computer-readable recording media.
Additional benefits and advantages of the disclosed embodiments will be apparent from the Specification and Drawings. The benefits and/or advantages may be individually obtained by the various embodiments and features of the Specification and Drawings, which need not all be provided in order to obtain one or more of such benefits and/or advantages.
Advantageous Effects of InventionOne or more exemplary embodiments or features disclosed herein provide a vehicle counting device and a vehicle counting method which accurately determine the number of vehicles, using vehicle sounds from other vehicles around a user vehicle. Subsequently, the vehicle counting device and the vehicle counting method notify a driver or the like of the user vehicle of the determined number of the vehicles, to support safe driving.
These and other advantages and features will become apparent from the following description thereof taken in conjunction with the accompanying Drawings, by way of non-limiting examples of embodiments of the present disclosure.
In the Drawings:
According to an exemplary embodiment disclosed herein, a vehicle counting device determines the number of nearby vehicles, using one or more vehicle sounds, the vehicle counting device including: a frequency analysis unit that analyzes a frequency of a surrounding sound that includes the one or more vehicle sounds and is detected by a vehicle sound detection microphone; a vehicle sound candidate selection unit that selects, as one or more vehicle sound candidates, one or more sounds included in the surrounding sound, based on a result of the analysis by the frequency analysis unit, each of the one or more sounds being in a frequency band where a sound pressure is greater than or equal to a predetermined threshold value; a phase curve calculation unit that calculates, for each of the selected one or more vehicle sound candidates, a phase curve indicating a time dependency of a phase; and a vehicle count determination unit that classifies the calculated one or more phase curves into at least one group, based on shapes of the calculated one or more phase curves, and determines a total number of the at least one group as the number of the nearby vehicles. In this way, it is possible to determine the number of the nearby vehicles, based on the shapes of the phase curves corresponding to the one or more vehicle sounds, that is, properties unique to the vehicle sounds. Thus, it is possible to accurately detect the number of vehicles around a user vehicle, using the one or more vehicle sounds.
For example, the phase curve calculation unit may calculate, for each of the selected one or more vehicle sound candidates, a quadratic approximate curve as the phase curve, based on a phase of a signal in each of time periods, the signal indicating the vehicle sound candidate. Here, the number of vehicles determination unit may classify the calculated one or more phase curves into the at least one group, based on a degree of similarity among quadratic coefficients of the one or more calculated quadratic approximate curves. In this way, it is possible to first determine a degree of similarity among acceleration and deceleration states of vehicles, and then determine the number of the vehicles even when plural vehicle sound candidates are present, because the number of the vehicles is determined by using the degree of similarity among the convex directions of the phase curves, and it is also possible to more accurately determine the number of the vehicles even in a complex situation such as a situation where plural vehicles are approaching.
For instance, the vehicle counting device may further include a vehicle sound extraction unit that calculates an error between a phase resulting from the analysis by the frequency analysis unit and a phase on the phase curve calculated by the phase curve calculation unit, and extracts, from regions resulting from the analysis by the frequency analysis unit, a region corresponding to a vehicle sound, based on the calculated error, wherein the number of vehicles determination unit may determine the number of the nearby vehicles, using a phase curve in the extracted region corresponding to the vehicle sound. In this way, it is possible to more accurately determine the number of the vehicles by using regions in a frequency spectrum each of which indicates a time dependency of a phase that is similar to a moving state (constant speed, acceleration, or deceleration) of a vehicle, that is, using only regions corresponding not to noises such as wind but to the vehicle sounds.
For example, the vehicle counting device may further include a notification unit that performs notification of the determined number of the nearby vehicles. Here, the notification unit may perform the notification in different modes depending on whether the determined number of the nearby vehicles is one or plural, or may perform the notification by sound. In this way, it is possible to notify the driver or the like of the number of the vehicles, thereby ensuring safe driving.
It is to be noted that these general and specific aspects are not only implemented using the vehicle counting device including such characteristic processing units but also may be implemented using a vehicle counting method including, as steps, the characteristic processing units of the vehicle counting device, and a computer program causing a computer to execute the characteristic steps of the vehicle counting method. It goes without saying that such a program can be distributed via a nonvolatile recording medium such as a Compact Disc-Read Only Memory (CD-ROM) or a communication network such as the Internet.
Hereinafter, certain exemplary embodiments are described in greater detail with reference to the accompanying Drawings. Each of the exemplary embodiments described below shows a general or specific example. The numerical values, shapes, structural elements, the arrangement and connection of the structural elements, steps, the processing order of the steps etc. shown in the following exemplary embodiments are mere examples, and therefore do not limit the inventive concept, the scope of which is defined in the appended Claims and their equivalents. Therefore, among the structural elements in the following exemplary embodiments, structural elements not recited in any one of the independent claims defining the most generic part of the inventive concept are described as arbitrary structural elements.
Embodiment 1The following describes a vehicle counting device according to Embodiment 1.
The vehicle sound detection microphone 101 is a microphone which detects a surrounding sound including vehicle sounds, that is, a microphone which detects at least vehicle sounds emanating from other vehicles such as engine sounds, motor sounds, and moving sounds, and outputs sound information (e.g., sound data in WAV format as shown by
The frequency analysis unit 102 performs a frequency analysis on the surrounding sound detected by the vehicle sound detection microphone 101. For instance, the frequency analysis unit 102 performs the Fourier transform on the surrounding sound, to obtain a frequency signal, an amplitude, a phase, and the like of the surrounding sound. It is to be noted that the Fourier transform performed by the frequency analysis unit 102 may be frequency transform using another frequency transform method such as a fast Fourier transform and the discrete cosine transform. More specifically, the frequency analysis unit 102 analyzes, for each analysis section that is a combination of one of predetermined time periods and one of predetermined frequency bands, sound pressure of the surrounding sound detected by the vehicle sound detection microphone 101, based on the surrounding sound.
Each of (a), (b), and (c) of
In an engine, a predetermined number of cylinders make piston motion to cause revolutions to a power train. An engine sound emanating from a vehicle includes: a sound dependent on engine revolutions; and a fixed vibration sound and an aperiodic sound that are independent of the engine revolutions. In particular, a main sound detectable from the outside of the vehicle is a periodic sound dependent on the engine revolutions. In this embodiment, attention is focused on the periodic sound dependent on the engine revolutions. In addition, an engine structure and engine revolutions differ from vehicle to vehicle, and thus a timbre differs. When plural vehicles are present, plural frequencies have a great magnitude of power in a spectrogram.
(a) of
In contrast, each of (b) and (c) of
The vehicle sound candidate selection unit 103 selects, as a vehicle sound candidate, a sound in a frequency band having sound pressure greater than or equal to a predetermined threshold value, from the surrounding sound, based on the analysis by the frequency analysis unit 102. In this embodiment, for instance, the vehicle sound candidate selection unit 103 selects, as the vehicle sound candidate, a frequency signal greater than or equal to a predetermined sound pressure threshold value for distinguishing between a vehicle sound and a noise, using the predetermined sound pressure threshold value. The predetermined sound pressure threshold value is, for example, −48 dB. Here, when a frequency spectrum is searched for a peak by using a peak search method or the like, and the number of peaks greater than or equal to a predetermined threshold value (−48 dB) is determined, it is possible to determine the number of the peaks as the number of vehicles. There is a case where a detectable vehicle sound is searched for a peak, using only the frequency having the greatest power such as a fundamental frequency and a predetermined frequency, due to the influence of surrounding noise or background noise, especially in the actual environment. Thus, as stated above, the number of the peaks greater than or equal to the predetermined threshold value may be determined as the number of the vehicles, using the predetermined threshold value. Alternatively, a noise level may be determined in advance by a spectrum subtraction method, and a remaining portion may be used as a vehicle sound.
There is, however, a case where even when one vehicle is present, peaks appear in frequency bands in a frequency spectrum actually obtained.
For instance, (a) of
In addition, when vehicles are present, it is more difficult to determine whether or not the vehicles are present, based only on a frequency band or power.
For example, (b) of
Especially, in a situation where vehicles are present and approaching, when it is wrongly determined that one vehicle is present, and a driver of a user vehicle is notified of only the presence of the one vehicle but not of the presence of the other approaching vehicles, there is a high possibility that the driver enters an intersection without worry after the one vehicle passes in front of the user vehicle, and subsequently the user vehicle collides with any of the other vehicles that are not detected. Thus, it is necessary to accurately determine the number of vehicles, and notify the determined number of the vehicles.
Furthermore, there is a possibility that a vehicle sound is canceled out by a noise of the wind, reflection, diffraction, and so on in the actual environment. In addition, in vehicle detection, it is necessary to determine instantaneously the presence or absence of a vehicle, and notify the presence of the vehicle.
In view of the above, in this embodiment, the vehicle counting device 100 determines the number of vehicles, using not only sound pressure (power) but also a phase, among information resulting from the frequency analysis by the frequency analysis unit 102. To do so, the vehicle counting device 100 according to this embodiment includes the phase curve calculation unit 104. The phase curve calculation unit 104 calculates, for each vehicle sound candidate selected by the vehicle sound candidate selection unit 103, a phase curve showing a time dependency of a phase.
Here, the term “phase” used in the present disclosure is defined with reference to
Moreover, (b) of
(c) of
It is to be noted that, in the sound signal processing, the fast Fourier transform (FFT), and so on, it is common to perform the convolution operation while the base waveform is being shifted along the temporal axis. When the convolution operation is performed while the base waveform is being shifted along the temporal axis, it is possible to correct a phase resulting from the convolution operation, to convert the phase into the phase defined in the present disclosure. The following describes the correction and conversion with reference to the drawings. In this embodiment, the base waveform is fixed without being shifted along the temporal axis, and a phase resulting from the convolution operation performed on the base waveform and the received engine sound is also referred to as a corrected phase.
Moreover, (b) of
(c) of
In this embodiment, the phase shown by
As shown by dotted circles 501 to 503 in
Here, when attention is focused on the variation of the frequency indicated by the streak (the peak) in the spectrogram, the frequency seldom varies randomly and is seldom discrete. The frequency clearly fluctuates in a predetermined manner at predetermined time intervals. For instance, the frequency indicated by the streak (the peak) clearly decreases in the section A shown by
The following describes a relationship between the decrease or increase in RPM and the phase of the engine sound. (a) of
It is to be noted that when a target sound has a certain frequency and the frequency of the base waveform is lower than that of the target sound, a phase is delayed gradually. Since, however, an amount of decrease is constant, the phase decreases linearly. In contrast, when a target sound has a certain frequency and the frequency of the base waveform is higher than that of the target sound, a phase is advanced gradually. Since, however, an amount of increase is constant, the phase increases linearly.
(a) of
(a) of
Thus, as shown by (d) of
Moreover, in this embodiment, it is possible to accurately determine the number of vehicles which cannot be accurately determined based on spectral power, by using a property that the phase significantly varies in a short time.
Furthermore, even when vehicle sounds are detected only instantaneously due to noises in the actual environment, it is possible to instantly determine the number of vehicles, using data obtained in the short time. Thus, it is possible to notify the driver of the accurate number of nearby vehicles in the short time.
In view of the above, the phase curve calculation unit 104 shown by
The following describes processing performed by the phase curve calculation unit 104 on the frequency band (i.e., the vehicle sound candidate) selected in the threshold value process performed by the vehicle sound candidate selection unit 103. In addition, here, the description is based on an example where the center frequency of the frequency band and the frequency of a base waveform match. To put it differently, it is determined whether or not a frequency f having a phase (here, a corrected phase ψ′(t) (=mod 2nψ(t)−2nft)) increases relative to an analysis frequency f (the frequency of the base waveform). It is to be noted that in this embodiment, the frequency analysis unit 102 regards the base waveform shown by
(b) of
A real part and an imaginary part of the frequency signal are expressed as x(t) and y(t), respectively. A phase ψ(t) and magnitude (power) P(t) of the frequency signal are expressed respectively as below.
[Math. 1]
ψ(t)=mod 2π(arctan(y(t)/x(t))) (Equation 1)
and
[Math. 2]
P(t)=√{square root over (x(t)2+y(t)2)}{square root over (x(t)2+y(t)2)} (Equation 2)
Here, the symbol t represents a time in the frequency signal.
(c) of
(d) of
In view of the above, in this embodiment, the vehicle count determination unit 105 focuses on phases of engine sounds, classifies, for each vehicle, the phases (i.e., phase curves) into at least one group based on time dependencies of phases (i.e., the shapes of the phase curves), and determines the number of vehicles.
A relationship between the increase and decrease in RPM and the time dependency of the phase can be expressed by the following equation.
[Math. 3]
ψ(t)=2π∫f(t)dt (Equation 3)
As shown by
[Math. 4]
f(t)=At+f0 (Equation 4)
Specifically, in a predetermined time period, the frequency f(t) at the time t can be linearly approximated using a line segment which increases or decreases from an initial value f0 in proportion to the time t (a proportionality coefficient A).
When the frequency f(t) is expressed by Equation 4, the phase ψ(t) at the time t is expressed as follows.
[Math. 5]
ψ(t)=2π∫f(t)dt=2π∫(At+f0)dt=πA2+2πf0t+ψ0 (Equation 5)
Here, ψ0 in the third term on the right-hand side represents an initial phase, and the second term (2nf0t) indicates that a phase is advanced by an angular frequency 2nf0t in proportion to the time t. The first term (nAt2) indicates that the phase can be approximated by a quadratic curve. Thus, in this embodiment, the phase curve calculation unit 104 calculates, for each vehicle sound candidate selected by the vehicle sound candidate selection unit 103, a quadratic curve as a phase curve by approximation, based on a phase of a signal at each time period, which indicates the vehicle sound candidate.
The following describes a method of calculating a corrected phase from a phase (a phase before correction) calculated while a base waveform is being shifted on a temporal axis.
It is to be noted that when the phase is calculated while the base waveform is being shifted on the temporal axis, as shown by (c) and (d) of
First, a reference time is determined. (a) of
Next, the phase curve calculation unit 104 determines times in frequency signals for which phases are corrected. In this example, times (t1, t2, t3, t4, and t5) indicated by five white circles in (a) of
Here, a phase of a frequency signal at the reference time t0 is expressed as follows.
[Math. 6]
ψ(t0)=mod 2π(arctan(y(t0)x(t0))) (Equation 6)
A phase of a frequency signal at each of the five times when the phases are corrected is expressed as follows.
[Math. 7]
ψ(ti)=mod 2π(arctan(y(ti)/x(ti))) (i=1,2,3,4,5) (Equation 7)
Each of the phases before correction is indicated by a cross mark in (a) of
[Math. 8]
P(ti)=√{square root over (x(ti)2+y(ti)2)}{square root over (x(ti)2+y(ti)2)} (i=1,2,3,4,5) (Equation 8)
Next,
[Math. 9]
ψ′(ti) (i=0,1,2,3,4,5)
In (b) of
[Math. 10]
Δψ=2πf(t2−t0) (Equation 9)
In view of the above, in (a) of
[Math. 11]
ψ′(t0)=ψ(t0) (Equation 10)
[Math. 12]
ψ′(ti)=mod 2π(ψ(ti)−2πf(ti−t0)) (i=1,2,3,4,5) (Equation 11)
In (b) of
The phase curve calculation unit 104 calculates a time dependencies of phases (corrected phases) as a curve (a phase curve that is a quadratic curve). First, the phase curve calculation unit 104 selects frequency signals to be used for calculating a phase shape. Assuming that a time when an analysis is performed is represented by t0, the phase curve calculation unit 104 calculates the phase shape from phases of frequency signals at the times t0 to t5. Here, the frequency signals used for calculating the phase curve are greater than or equal to a predetermined value in number (in this example, the six frequency signals at the times t0 to t5). This is because it is possible to avoid the difficulty in determining the regularity of the time dependencies of the phases when the number of frequency signals selected for calculating a phase distance is small. Predetermined duration may be determined based on the characteristics of a time dependency of a phase in an engine sound. An analysis section may be changed depending on an area where a vehicle is present. For example, the analysis section is reduced, because a phase variation is rapid in a narrow street area or an intersection area where a vehicle often accelerates or decelerates, and the analysis section is extended for an area where the vehicle moves at a relatively constant speed. In addition, the analysis section may be changed depending on a frequency band to be analyzed. The analysis section suitable for the frequency band may be used appropriately. For instance, the analysis section is extended in the case of a low frequency such as 200 Hz or less, and the analysis section is reduced for a relatively high frequency band such as 200 Hz or greater.
The phase curve calculation unit 104 calculates, by approximation, the phase curve from the phases of the selected frequency signals. The phase curve is approximated according to, for example, the following quadratic polynomial (Equation 12 below).
[Math. 13]
Ψ(t)=A2t2+A1t+A0 (Equation 12)
It is to be noted that items on the right-hand side of respective Equations 13 and 14 are expressed as follows.
The vehicle count determination unit 105 classifies, into at least one group, the phase curves that are the quadratic curves calculated by the phase curve calculation unit 104, based on the shapes of the phase curves (more specifically, by using a degree of similarity among the quadratic coefficients of the quadratic curves (the uniformity of signs)), and determines the number of the obtained groups as the number of vehicles.
Specifically, the vehicle count determination unit 105 distinguishes, by using the uniformity of convex directions of the phase curves calculated by the phase curve calculation unit 104, vehicles, to accurately determine the number of the vehicles. When the coefficient A2 calculated according to Equation 12 is positive, that is, a phase curve is convex downward, the vehicle is considered to be accelerating due to the increase in RPM. In contrast, when the coefficient A2 is negative, that is, a phase curve is convex upward, the vehicle is considered to be decelerating due to the decrease in RPM. Suppose that two engine sounds (i.e., two streaks (peaks) at different frequencies) are detected and that convex directions of phase curves differ from each other for the two engine sounds, the vehicle count determination unit 105 recognizes the two engine sounds as being emanated from different vehicles, and determines that two vehicles have been present. On the other hand, suppose that the convex directions of the phase curves are identical to each other for the two engine sounds, the vehicle count determination unit 105 recognizes the two engine sounds as being emanated from the same vehicle, and determines that one vehicle has been present. As stated above, the vehicle count determination unit 105 classifies, into the at least one group, the phase curves that are the quadratic curves calculated by the phase curve calculation unit 104, using the degree of similarity among the quadratic coefficients of the quadratic curves (the uniformity of signs), and determines the number of the obtained groups as the number of the vehicles.
In contrast, similar to (b) of
An analysis region 1 is at the frequency of 125 Hz and has the duration of 100 ms between 100 ms and 200 ms, an analysis region 2 is at the frequency of 85 Hz and has the duration of 100 ms between 100 ms and 200 ms, an analysis region 3 is at the frequency of 65 Hz and has the duration of 100 ms between 100 ms and 200 ms, and an analysis region 4 is at the frequency of 40 Hz and has the duration of 100 ms between 100 ms and 200 ms. (To put it differently, the analysis regions 1 to 4 are regions selected by the vehicle sound candidate selection unit 103.) Respective phases in the analysis regions 1 to 4 are represented by phase curves indicated by directional lines. The phase curves are convex downward in the analysis regions 1 and 3, and the vehicle count determination unit 105 classifies the two streaks (peaks) corresponding to the analysis regions 1 and 3 as one vehicle. In addition, the phase curves are convex upward in the analysis regions 2 and 4, and the vehicle count determination unit 105 classifies the two streaks (peaks) corresponding to the analysis regions 2 and 4 as another vehicle. Thus, the vehicle count determination unit 105 determines that the two vehicles in total have been present.
The notification unit 106 is, for example, a display control unit or a display unit which notifies the driver of the number of the vehicles determined by the vehicle count determination unit 105. In addition, the notification unit 106 controls a notification mode depending on the number of the vehicles. For instance, when the vehicle counting device is included in a car navigation system or the like, an output mode is set to sound, and the notification unit 106 gives an audio alert about the approach of vehicle and the number of vehicles.
(a) of
(b) of
When the approach of vehicle is notified, it is necessary to notify the driver as quickly and simply as possible. The notification by the “beep” by the notification unit 106 produces an effect of urging the driver to drive safely. Moreover, the mere notification of the approach of vehicle does not allow the driver to know whether one or more vehicles are approaching, and the driver gets confused accordingly. In view of the above, when the one vehicle is determined to be present, the notification unit 106 beeps once, and when the two vehicles are determined to be present, the notification unit 106 beeps twice, thereby notifying the driver of the difference. In other words, the notification unit 106 notifies the driver in a different mode depending on whether the vehicle count determination unit 105 determines that the one vehicle has been present or determines that the two vehicles have been present.
It is to be noted that the cases where the one nearby vehicle is present and where the two nearby vehicles are present are described as the example in this embodiment, the present disclosure is not limited to the cases. When plural vehicles such as three vehicles are present, the notification unit 106 may beep as many times as the number of the vehicles.
Furthermore, when two or more vehicles such as three vehicles are present, the notification unit 106 may notify that the plural vehicles are present, by beeping twice, for instance, as in the case where the two vehicles are present, instead of as many times as the number of the vehicles, and may switch between notifications depending on whether the one or more vehicles are determined to be present. Even notifying the driver of whether the one or more vehicles are determined to be present often leads to assistance in making a decision by the driver. The driver can make a decision such as a decision to enter the intersection immediately after the one vehicle passes, when the only one vehicle is determined to be present, and a decision to wait when the plural vehicles are determined to be present, whether the number of the plural vehicles is two or three. A procedure described in this embodiment may be used to distinguish whether the one or more vehicles are determined to be present, and the notification unit 106 may switch between the notification modes. Moreover, although the output mode is described as being set to the sound in this embodiment, the notification unit 106 in this embodiment is not limited to use the sound. When notifying the presence of vehicle visually, the notification unit 106 may display as many symbols of vehicles as the number of vehicles.
As stated above, the notification unit 106 may change control depending on whether the one or more vehicles are determined to be present. It may be important for the driver to know whether the one or more vehicles are determined to be present, depending on a driving situation. For instance, suppose that the driver would like to know the presence of approaching vehicle at the intersection, when one vehicle is determined to be present, the user vehicle can enter the intersection after the one vehicle passes, but when plural vehicles are determined to be present, the user vehicle needs to wait whether the number of the plural vehicles is two or three. Here, the notification modes may be switched between depending on whether the one or more vehicles are determined to be present.
The following describes an operation flow of the vehicle counting device 100 according to this embodiment with reference to flow charts shown by
First, the vehicle sound detection microphone 101 detects one or more vehicle sounds (step S101). Next, the frequency analysis unit 102 performs a frequency analysis on a surrounding sound (step S102).
Then, the vehicle sound candidate selection unit 103 selects, from the surrounding sound on which the frequency analysis is performed, predetermined frequency bands and time periods as vehicle sound candidates (step S103). For example, the vehicle sound candidate selection unit 103 sets a threshold value for sound pressure, and selects frequency bands and time periods greater than or equal to the predetermined threshold value as vehicle sound candidates. Next, the phase curve calculation unit 104 calculates, as a phase curve, a phase shape in each of the selected periods (vehicle sound candidates) which varies with time (step S104). In other words, the phase curve calculation unit 104 calculates the phase curve from a phase (corrected phase) in each of the periods (vehicle sound candidates) selected by the vehicle sound candidate selection unit 103, according to Equations 13 to 15.
Then, the vehicle count determination unit 105 determines the number of vehicles based on the calculated phase curves (step S105).
In contrast, when the phase curve projects downward (is convex downward), the vehicle count determination unit 105 sets a downward flag on (step S204). Finally, the vehicle count determination unit 105 calculates the number of vehicles by referring to the number of flags (step S205). As above, the vehicle count determination unit 105 determines the number of the vehicles by using the degree of similarity among the quadratic coefficients of the phase curves that are the quadratic curves calculated by the phase curve calculation unit 104.
It is to be noted that the binarization process (the steps S202 to S204) is performed in this embodiment, a vehicle counting method is not limited to this. For example, as in another detailed flow of the step S105 shown by
Finally, the notification unit 106 notifies the determined number of the vehicles (step S106).
As described above, the vehicle counting device 100 according to this embodiment accurately determines the number of the vehicles based on the shapes of the phase curves corresponding to the vehicle sounds, that is, the properties unique to the vehicle sounds, and notifies the driver of the determined number of the vehicles, thereby supporting the safe driving. In addition, although, when the approach of vehicle is notified, it is necessary to notify the driver as quickly and simply as possible, the vehicle counting device 100 according to this embodiment determines the number of the vehicles based on the time dependencies of the phases, that is, determines the number of the vehicles in a short time such as several hundreds ms.
Moreover, the mere notification of the approach of vehicle as in the conventional technique does not allow the driver to know whether the one or more vehicles are approaching, and the driver gets confused accordingly. In contrast, the vehicle counting device 100 according to this embodiment determines the number of the vehicles based on the shapes of the phase curves unique to the vehicle sounds, and thus the vehicle counting device 100 is more effective when the plural vehicles are present and avoids a danger that the user vehicle enters the intersection mistakenly after the one vehicle passes.
As above, in this embodiment, the vehicle counting device 100 determines the number of the vehicles, using the shapes of the phase curves corresponding to the vehicle sounds at the same time. Specifically, as shown by (a) of
It is to be noted that the vehicle counting method is not limited to the above. For instance, the number of vehicles may be determined by using shapes of phase curves, each of which has predetermined duration, within the predetermined duration.
Generally speaking, a vehicle is less likely to frequently repeat acceleration and deceleration in a short time, and may keep accelerating or decelerating during a certain period of time. Thus, shapes of phases in the period of time are often similar to each other. In addition, an engine sound in the actual environment does not necessarily include sound pressure having a predetermined threshold value due to the influence of the surrounding noise or the like, and there is a case where all vehicle sounds cannot be detected at the same time. Thus, the number of the vehicles may be determined based not on the shapes of the phases at the same time but on the shapes of the phase curves with the predetermined duration. Consequently, it is possible to determine the number of the vehicles in a manner more suitable for the actual environment.
Moreover, the shapes of the phase curves may be determined using mutually different analysis regions in the same streak (peak) in the spectrogram. (b) of
As above, the vehicle is less likely to frequently repeat the acceleration and deceleration in a short time, and the shapes of the phase curves such as especially the convex directions of the phase curves are often similar to each other within the predetermined duration. However, the phase of the engine sound in the actual environment is disturbed due to the influence of the surrounding noise or the like, and the resultant shape of the phase contains an error. In view of the above, it is possible to determine, with a higher degree of reliability, the shapes of the phase curves by determining states of the vehicles such as the acceleration of the vehicles based not on the shapes of the phase curves at a certain moment but on the shapes of the phase curves within the predetermined duration. As a result, it is also possible to increase the accuracy of determining the number of vehicles.
Moreover, although the vehicle sounds are classified into the groups based on the convex directions of the phase curves and the number of the vehicles is determined in this embodiment, the vehicle counting method is not limited to this. When plural vehicles are present, phase curves corresponding to vehicle sounds emanated from even the vehicles similarly accelerating have different degrees of convex (the quadratic coefficients A2 in Equation 12). Here, the phase curves may be classified into groups based on the degrees of convex (the quadratic coefficients A2 in Equation 12), and the number of vehicles may be determined.
Furthermore, although the shapes of the phase curves that are the quadratic curves are classified into the groups based on the convex directions of the quadratic curves (the degree of similarity among the quadratic coefficients of the quadratic curves (the uniformity of signs)) and the number of the vehicles is determined in this embodiment, the phase curve classification method is not limited to this. For example, a predetermined shape of a phase curve may be stored in advance, and shapes of phase curves may be classified into groups based on an error between the predetermined shape and each of the shapes. Specifically, a phase curve which is convex upward and has a certain curvature (hereinafter, referred to as a phase curve 1) and a phase curve which is convex downward and has a certain curvature (hereinafter, referred to as a phase curve 2) are stored in advance. An error between a phase at each time obtained from a detected vehicle sound and a phase on each phase curve is calculated, and a phase curve having a small error is used as a phase curve corresponding to the vehicle sound. Phase curves having a small error from the phase curve 1 may be classified into a group, and phase curves having a small error from the phase curve 2 may be classified into another group, thereby determining the number of the vehicles. This eliminates the need for calculating an approximate curve, resulting in reducing a processing load, and also produces an effect on an in-vehicle application required to instantly determine the number of the vehicles and notify the driver of the determined number of the vehicles.
Moreover, the number of vehicles may not be determined accurately, and it may be determined whether one or more vehicles are present. This produces the effect of the safe driving support for preventing the driver from being misled by mistakenly notifying the driver of the presence of the one vehicle even though the more vehicles are present.
The herein disclosed subject matter is to be considered descriptive and illustrative only, and the appended Claims are of a scope intended to cover and encompass not only the particular embodiment disclosed, but also equivalent structures, methods, and/or uses.
Moreover, although the vehicle sounds (phase curves) are classified into the groups based on the convex directions of the phase curves and the number of the vehicles is determined in this embodiment, the vehicle counting method is not limited to this. Each of the convex directions of the phase curves is considered as information indicating the acceleration and deceleration of a vehicle. Here, not only the vehicle sounds may be classified into the groups but also it may be determined whether or not the vehicles are accelerating or decelerating, and the notification may be switched based on the acceleration and deceleration.
For instance, as shown by (b) of
The following describes a vehicle counting device according to Embodiment 2.
In Embodiment 1, the vehicle counting device performs the frequency analysis on the vehicle sounds, calculates the phase curves from the phases resulting from the frequency analysis, and determines the number of the vehicles based on the phase curves. In contrast, in this embodiment, the vehicle counting device extracts one or more vehicle sounds from a mixed sound including vehicle sounds and noises such as a noise of the wind, based on phase curves, and determines the number of vehicles based on the extracted one or more vehicle sounds.
The vehicle sound extraction unit 107 extracts one or more vehicle sounds based on the phase curves calculated by the phase curve calculation unit 104. In other words, the vehicle sound extraction unit 107 calculates an error between a phase obtained from a result of the frequency analysis by the frequency analysis unit 102 and a phase on each of the phase curves calculated by the phase curve calculation unit 104, and extracts regions corresponding to the vehicle sounds from regions obtained from the result of the frequency analysis by the frequency analysis unit 102, based on the calculated error.
Then, the vehicle count determination unit 105 determines the number of the vehicles, using the phase curves in the regions corresponding to the vehicle sounds which are extracted by the vehicle sound extraction unit 107.
When, for instance, installed in the user vehicle and used in the actual environment, the vehicle counting device according to this embodiment is significantly influenced by a noise such as a surrounding environmental sound and a wind noise produced by the moving user vehicle.
In the case of the example shown by (b) of
In contrast, (c) of
Furthermore, (d) of
In contrast, (e) of
It is to be noted that in this embodiment, the vehicle count determination unit 105 uses the method described in Embodiment 1 to determine the number of the vehicles, using the regions extracted by the vehicle sound extraction unit 107.
The following describes operations of the vehicle counting device 200 according to this embodiment with reference to flow charts shown by
The same reference signs are assigned to the same steps as those in Embodiment 1, and descriptions thereof are omitted.
A step (step S401) is added in which the vehicle sound extraction unit 107 extracts vehicle sounds after the phase curve calculation unit 104 calculates the phase curves as described in Embodiment 1 (step S104).
In the step S401, first, the vehicle sound extraction unit 107 calculates an error between a phase at each time which is obtained from the result of the frequency analysis by the frequency analysis unit 102 and each of the phase curves calculated by the phase curve calculation unit 104 (step S501), and calculates an average of the errors (step S502). Then, the vehicle sound extraction unit 107 determines whether or not the average is less than a predetermined threshold value (e.g., 20 degrees) (step S503), and extracts, when the average is determined to be less than the threshold value (Yes in step S503), one or more regions in a spectrogram corresponding to each time as one or more vehicle sounds (step S504). Subsequently, the vehicle count determination unit 105 uses the method described in Embodiment 1 to determine the number of vehicles, using the one or more regions extracted by the vehicle sound extraction unit 107 (step S105). Finally, the notification unit 106 performs a notification according to the number of the vehicles determined by the vehicle count determination unit 105 (step S106).
As described above, the vehicle counting device 200 according to this embodiment extracts the one or more vehicle sounds from the mixed sound based on the phase curves, determines the number of the vehicles based on the extracted one or more vehicle sounds, and notifies the driver of the number of the vehicles. In addition, the method described in this embodiment makes it possible to accurately extract the one or more vehicle sounds even in an environment with noises such as a wind noise produced by a moving vehicle, to determine the number of the vehicles based on the one or more vehicle sounds, and to notify the number of the vehicles.
Although the vehicle counting device is described in the embodiments and modifications disclosed herein, the present disclosure is not limited to the embodiments and modifications.
For instance, any of the elements in the embodiments and modifications disclosed herein may be arbitrarily combined within the scope of the present disclosure.
Moreover, it should be considered that the embodiments and modifications disclosed herein are exemplary in all respects and not restrictive at all. It is intended that the scope of the present disclosure is indicated not by the above description of the embodiments but by the claims, and that all changes that have equivalent meaning as and fall within the claims are included in the scope of the present disclosure.
Moreover, the elements included in the vehicle counting device according to the embodiments may be implemented as hardware such as a dedicated electronic circuit, or may be configured as a computer system including a microprocessor, a ROM, a RAM, a hard disk drive, a display unit, a keyboard, a mouse, and so on. A computer program is stored in the RAM or the hard disk drive. Each of systems or devices achieves its functions as a result of the microprocessor operating according to the computer program. Here, in order to achieve predetermined functions, the computer program is configured by combining a plurality of instruction codes indicating instructions for a computer.
Furthermore, part or all of the elements included in the vehicle counting device according to the embodiments and modifications may be configured from one system LSI (Large Scale Integration). The system LSI is a super-multifunctional LSI manufactured with a plurality of components integrated on a single chip, and specifically is a computer system including a microprocessor, a ROM, and a RAM, for example. A computer program is stored in the RAM. The system LSI achieves its functions as a result of the microprocessor operating according to the computer program.
Still furthermore, part or all of the elements included in the vehicle counting device according to the embodiments and modifications may be configured from a IC card detachable from each system or each device or a stand-alone module. The IC card or the module is a computer system including a microprocessor, a ROM, a RAM, and so on. The IC card or the module may include the super-multifunctional LSI. The IC card or the module achieves its functions as a result of the microprocessor operating according to a computer program. The IC card or the module may have tamper-resistance.
Moreover, the vehicle counting device in the present disclosure may be realized as the above-described method (the vehicle counting method). In addition, the vehicle counting device in the present disclosure may be a computer program realizing these methods with a computer, or a digital signal of the computer program.
Furthermore, the vehicle counting device and the vehicle counting method in the present disclosure may be realized as a non-transitory computer-readable recording medium on which the computer program or the digital signal is recorded, such as a flexible disk, a hard disk, a CD-ROM, an MO, a DVD, a DVD-ROM, a DVD-RAM, a BD (Blu-ray Disc (registered trademark)), and a semiconductor memory. In addition, the vehicle counting device and the vehicle counting method in the present disclosure may be the digital signal recorded on the non-transitory computer-readable recording medium.
Moreover, the computer program or the digital signal may be transmitted via, for instance, an electric telecommunication line, a wireless or wired communication line, a network such as the Internet, or data broadcasting.
Furthermore, the vehicle counting device and the vehicle counting method in the present disclosure may be a computer system including a microprocessor and a memory storing the computer program, wherein the memory stores the computer program and the microprocessor operates according to the computer program.
Moreover, another independent computer system may execute the computer program or the digital signal that is recorded on the non-transitory computer-readable recording medium and transmitted to the other independent computer system or that is transmitted via the network or the like to the other independent computer system.
INDUSTRIAL APPLICABILITYOne or more exemplary embodiments disclosed herein are applicable to devices that determine the number of vehicles, and particularly to a device that detects the presence or absence of any vehicle and the number of vehicles, using one or more vehicle sounds from at least one vehicle around a user vehicle.
Claims
1. A vehicle counting device that determines the number of nearby vehicles, using one or more vehicle sounds, the vehicle count determination device comprising:
- a frequency analysis unit configured to analyze a frequency of a surrounding sound that includes the one or more vehicle sounds and is detected by a vehicle sound detection microphone;
- a vehicle sound candidate selection unit configured to select, as one or more vehicle sound candidates, one or more sounds included in the surrounding sound, based on a result of the analysis by the frequency analysis unit, each of the one or more sounds being in a frequency band where a sound pressure is greater than or equal to a predetermined threshold value;
- a phase curve calculation unit configured to calculate, for each of the selected one or more vehicle sound candidates, a phase curve indicating a time dependency of a phase; and
- a vehicle count determination unit configured to classify the calculated one or more phase curves into at least one group, based on shapes of the calculated one or more phase curves, and determine a total number of the at least one group as the number of the nearby vehicles.
2. The vehicle counting device according to claim 1,
- wherein the phase curve calculation unit is configured to calculate, for each of the selected one or more vehicle sound candidates, a quadratic approximate curve as the phase curve, based on a phase of a signal in each of time periods, the signal indicating the vehicle sound candidate.
3. The vehicle counting device according to claim 2,
- wherein the number of vehicles determination unit is configured to classify the calculated one or more phase curves into the at least one group, based on a degree of similarity among quadratic coefficients of the one or more calculated quadratic approximate curves.
4. The vehicle counting device according to claim 1, further comprising
- a vehicle sound extraction unit configured to calculate an error between a phase resulting from the analysis by the frequency analysis unit and a phase on the phase curve calculated by the phase curve calculation unit, and extract, from regions resulting from the analysis by the frequency analysis unit, a region corresponding to a vehicle sound, based on the calculated error,
- wherein the number of vehicles determination unit is configured to determine the number of the nearby vehicles, using a phase curve in the extracted region corresponding to the vehicle sound.
5. The vehicle counting device according to claim 1, further comprising
- a notification unit configured to perform notification of the determined number of the nearby vehicles.
6. The vehicle counting device according to claim 5,
- wherein the notification unit is configured to perform the notification in different modes depending on whether the determined number of the nearby vehicles is one or plural.
7. The vehicle counting device according to claim 5,
- wherein the notification unit is configured to perform the notification by sound.
8. A vehicle counting method for determining the number of nearby vehicles, using one or more vehicle sounds, the vehicle counting method comprising:
- analyzing a frequency of a surrounding sound including the one or more vehicle sounds;
- selecting, as one or more vehicle sound candidates, one or more sounds included in the surrounding sound, based on a result of the analysis in the analyzing, each of the one or more sounds being in a frequency band where a sound pressure is greater than or equal to a predetermined threshold value;
- calculating, for each of the selected one or more vehicle sound candidates, a phase curve indicating a time dependency of a phase; and
- classifying the calculated one or more phase curves into at least one group based on shapes of the calculated one or more phase curves, and determining a total number of the at least one group as the number of the nearby vehicles.
9. A non-transitory computer-readable recording medium for use in a computer, the recording medium having a computer program recorded thereon for causing the computer to execute the vehicle counting method according to claim 8, the computer program being for use by a vehicle counting device that determines the number of nearby vehicles, using one or more vehicle sounds.
Type: Application
Filed: Aug 28, 2012
Publication Date: Dec 20, 2012
Patent Grant number: 9205787
Inventors: Mototaka Yoshioka (Osaka), Shinichi Yoshizawa (Osaka)
Application Number: 13/596,535
International Classification: G06F 15/00 (20060101);